JOURNAL OF
BIOMEDICAL SEMANTICS
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2
h p://www.jbiomedsem.com/con en /4/1/2
SOFTWARE Open Access
OPPL-Galaxy, a Galaxy ool o enhancing
on ology exploi a ion as pa o bioin o ma ics
wo kflows
Mikel Ega˜
na A angu en1,2*,JesualdoTom
´
as Fe n´
andez-B eis3, Ch is Mungall4, E ick An ezana5,
Alejand o Rod ´
ıguez Gonz´
alez2and Ma k D Wilkinson2
Abs ac
Backg ound: Biomedical on ologies a e key elemen s o building up he Li e Sciences Seman ic Web. Reusing and
building biomedical on ologies equi es flexible and e sa ile ools o manipula e hem efficien ly, in pa icula o
en iching hei axioma ic con en . The On ology P e P ocesso Language (OPPL) is an OWL-based language o
au oma ing he changes o be pe o med in an on ology. OPPL augmen s he on ologis s’ oolbox by p o iding a
mo e efficien , and less e o -p one, mechanism o en iching a biomedical on ology han ha ob ained by a manual
ea men .
Resul s: We p esen OPPL-Galaxy, a w appe o using OPPL wi hin Galaxy. The unc ionali y deli e ed by OPPL
(i.e. au oma ed on ology manipula ion) can be combined wi h he ools and wo kflows de ised wi hin he Galaxy
amewo k, esul ing in an enhancemen o OPPL. Use cases a e p o ided in o de o demons a e OPPL-Galaxy’s
capabili y o en iching, modi ying and que ying biomedical on ologies.
Conclusions: Coupling OPPL-Galaxy wi h o he bioin o ma ics ools o he Galaxy amewo k esul s in a sys em ha
is mo e han he sum o i s pa s. OPPL-Galaxy opens a new dimension o analyses and exploi a ion o biomedical
on ologies, including au oma ed easoning, pa ing he way owa ds ad anced biological da a analyses.
Backg ound
Among he a ious s eps ha a ypical li e-sciences
esea ch cycle comp ises, in o ma ion ex ac ion om
aw da a (and i s dissemina ion o he communi y)
emains as one o he mos ele an ones. New biological
insigh s a e gene a ed by combining in o ma ion om di -
e en sou ces wi h he expe ise o scien is s. Ne e he-
less, in eg a ing in o ma ion and gene a ing knowledge
ou o i is s ill a challenging ask, as he in o ma ion is e-
quen ly cap u ed in compu a ionally opaque o ma s and
dispe sed o e he Web in esou ces wi h idiosync a ic
schemas.
The Seman ic Web [1] aims o o e come he issue o
compu a ionally opaque and dispe se in o ma ion in he
*Co espondence: [email p o ec ed]
1On ology Enginee ing G oup, School o Compu e Science, Technical
Uni e si y o Mad id (UPM), Boadilla del Mon e, 28660, Spain
2Biological In o ma ics G oup, Cen e o Plan Bio echnology and Genomics
(CBGP), Technical Uni e si y o Mad id (UPM), Pozuelo de Ala c´
on, 28223, Spain
Full lis o au ho in o ma ion is a ailable a he end o he a icle
Web wi h a se o echnologies and s anda ds defined by
he W3C: RDF [2], SPARQL [3] and OWL [4]. The e o e,
hese s anda ds a e inc easingly used by he Li e Sciences
communi y o in eg a e in o ma ion (RDF), o que y i
(SPARQL), and o axioma ically encode consensus knowl-
edge abou such in o ma ion in on ologies (OWL), in he
so-called Li e Sciences Seman ic Web [5].
Biomedical on ologies a e essen ial o he Li e Sciences
Seman ic Web since hey offe compu a ionally p ocess-
able and o en Web-o ien ed ep esen a ions o ag eed-
upon domain knowledge. The Gene On ology (GO) [6]
s ands ou as one o he mos in ensely cu a ed and
used biomedical on ologies; o he impo an biomedi-
cal on ologies can be ound a he Open Biological and
Biomedical On ologies Found y [7], a p ojec ha hos s
biomedical on ologies ha ollow ce ain design p in-
ciples ( eusabili y, o hogonali y, e c.). Addi ionally, he
Na ional Cen e o Biomedical On ology (NCBO) offe s
access o biomedical on ologies h ough BioPo al [8],
including a se o Web Se ices.
© 2013 Ega˜
na A angu en e al.; licensee BioMed Cen al L d. This is an Open Access a icle dis ibu ed unde he e ms o he
C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/2.0), which pe mi s un es ic ed use,
dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 2 o 16
h p://www.jbiomedsem.com/con en /4/1/2
Cu en biomedical on ologies suppo a b oad ange o
asks: axioma ically ich on ologies a e used o in ense
au oma ed easoning [9], axioma ically lean on ologies
ac as ocabula ies o Linked Da a [10], and ypically
o he unc ions in be ween [11]. In o de o ulfill such
unc ions, biomedical on ologies should be adap ed o
fi scien is s’ equi emen s, especially when eusing p e-
exis ing on ologies: addi ion o emo al o axioms and
en i ies, in e ence in ela ion o ex e nal on ologies, selec-
i e ma e ialisa ion o in e ed axioms, complex que ying,
and so o h.
Manipula ing biomedical on ologies can be a labo i-
ous ask since hey a e egula ly g owing in size [12]
and axioma ic complexi y [13]. The e o e, ad anced ools
a e needed o efficien ly pe o ming such manipula ion
[14]. The On ology P e P ocesso Language (OPPL) [15]
offe s he possibili y o au oma ing his kind o on ology
manipula ion. By using OPPL, he on ologis can define
he in ended manipula ion in an OPPL sc ip as a se ies
o addi ions o emo als o axioms o be pe o med in a
conc e e on ology. The e o e, he use o OPPL makes he
on ology manipula ion p ocess mo e efficien , sus ainable
and less e o -p one.
OPPL capabili ies ha e al eady been demons a ed: i
has been used o build an on ology ans o ma ion se ice
[16] and o applying [17-20] o de ec ing [21] On ol-
ogyDesignPa e ns(ODPs).Also,i ispa o Populous,
an applica ion o adding con en om sp eadshee s o
on ologies [22].
OPPL’s e sa ili y and unc ionali y canno be exploi ed
di ec ly wi hin he ypical bioin o ma ics analyses. Galaxy,
a Web se e o combining a ious genomic-o ien ed
ools in o wo kflows [23], offe s an ideal pla o m o
making OPPL pa o bioin o ma ics analyses. The e-
o e, we ha e de eloped OPPL-Galaxy, a ool o execu e
OPPL sc ip s om wi hin Galaxy. OPPL-Galaxy enhances
OPPL’s unc ionali y, i.e. au oma ed on ology manipula-
ion, by p o iding he possibili y o dynamically sending
OPPL’s ou pu , ha is, an imp o ed on ology, o o he
Galaxy ools (and making OPPL capable o consuming
on ologies as inpu om o he Galaxy ools).
This pape p esen s an o e iew o OPPL-Galaxy’s
design and implemen a ion, including es ed use cases
ha p o ide a basis o c ea ing mo e complex analy-
ses. OPPL-Galaxy is also compa ed o o he ools and i s
benefi s and limi a ions a e discussed.
Implemen a ion
OPPL
OPPL implemen s i s own syn ax: an ex ension o he
Manches e OWL Syn ax (MOS) [24] ha includes key-
wo ds like ADD ( o add an axiom), REMOVE ( o emo e an
axiom), SELECT ( o selec en i ies), and so on. An OPPL
sc ip defines a que y and some ac ions ha should be
pe o med agains he e ie ed en i ies (see ‘Basic usage’
use case in Sec ion Resul s). A que y can combine a i-
ables ( o be bound by a se o named en i ies) and ac ual
named en i ies o he a ge on ology (OWL classes, p op-
e ies, o indi iduals). An impo an cons ain in OPPL
specifies ha e e y a iable mus esol e o a g oup o
named en i ies (o none), no an anonymous OWL exp es-
sion, o ensu e ha que ies can be answe ed. The ollow-
ing ypes o que ies can be defined in OPPL (all he que ies
mix a iables wi h OWL exp essions):
•OWL que ies ha exploi au oma ed easoning.
•Syn ac ic OWL que ies ha only wo k wi h he
asse ed axioms.
•Que ies ha use a egula exp ession o ma ch
anno a ion alues like d s:label.
The ac ions a e based on he addi ion o emo al o
axioms o any complexi y o/ om en i ies e ie ed by
he que y (OWL classes, p ope ies, o ins ances). Once
an OPPL sc ip has been defined, he OPPL engine is
passed his sc ip and he on ology o be modified. The
OPPL engine, in u n, modifies he on ology acco ding o
he changes defined in he OPPL sc ip , gene a ing a new
on ology (Figu es 1 and 2).
Galaxy
Galaxy offe s an open, Web-based pla o m o pe o m-
ing genomic analyses [23]. In Galaxy se e al ools can
be combined, anging om simple da a manipula ions
(e.g. ex manipula ion) o complex analyses (e.g. s a is ical
analysis o Nex -Gene a ion Sequencing da a). Such a ool
o ches a ion can be execu ed om wi hin a single Web
in e ace: he ou pu o a ool can be sen o o he ools
as inpu , easing he cons uc ion o wo kflows by combin-
ing ecu en asks. Mo eo e , a his o y o all pe o med
ac ions is s o ed, so he analyses can be ep oduced a any
ime and sha ed wi h o he use s. Galaxy wo kflows can
be buil om he use s’ his o y and sha ed. Finally, he
wo kflows can be mig a ed o o he sys ems, like o he
Galaxy se e s o myExpe imen [25].
Apa om i s unc ionali y and ease o use, ano he
appealing ea u e o Galaxy is i s ex ensibili y, allowing
a s aigh o wa d in eg a ion o command-line ools: he
only equi emen is o c ea e an XML file con aining a
desc ip ion o he ool’s Web in e ace and inpu s/ou pu s
[26].
OPPL-Galaxy
OPPL can be execu ed h ough he g aphical in e ace o
P o ´
eg´
e [27] and Populous. Despi e hose possible means
o manipula ing on ologies, OPPL canno be used as
pa o a wo kflow, limi ing he possibili ies o including
o he bioin o ma ics analysis ools, unless a ailo ed Ja a
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 3 o 16
h p://www.jbiomedsem.com/con en /4/1/2
Figu e 1 Toy on ology o OWL ende ing con en ion. Toy on ology o illus a e he con en ion o ep esen ing abs ac OWL s uc u es in
Figu es depic ing use cases. Abo e, he on ology is ende ed using MOS; below, he on ology is ende ed wi h he same con en ion as in Figu es 2,
5, 6, 10 and 14. In hose Figu es, howe e , names o OWL en i ies a e no included in he on ologies, since OPPL sc ip s ac on abs ac s uc u es (any
axioma ic pa e n ha ma ches he que y). Solid ci cle: named class; do ed ci cle: anonymous class; do : named indi idual; solid a ow:
subClassO axiom; do ed a ow: iple ( ela ion be ween indi iduals); line ending in ci cle: es ic ion ( he small ci cle poin s o he fille class;
he e is no dis inc ion be ween necessa y and necessa y/sufficien condi ions)a.
p og am is w i en using he OPPL API. OPPL-Galaxy fills
ha gap by offe ing an enhanced e sion o OPPL ha
can be used in combina ion wi h o he Galaxy ools. To
ha end, an OPPL w appe was de eloped as a media o
be ween Galaxy and bo h he OPPL 2 API [28] and he
OWL API [29] (Figu e 3).
OPPL-Galaxy akes as inpu a a ge on ology and an
OPPL sc ip : bo h a e ac s a e uploaded o Galaxy by he
use o p oduced as ou pu by ano he Galaxy ool. I gen-
e a es a new on ology ha has been changed acco ding o
he ins uc ions defined in he OPPL sc ip , hus axioms
a e added o emo ed. The OPPL-Galaxy Web in e ace
p esen s he ollowing op ions (Figu e 4):
•Ta ge on ology: he inpu on ology ha will be
modified by he OPPL sc ip . Since OPPL-Galaxy
Figu e 2 OPPL pipeline. The OPPL engine akes an on ology (ci cle g oup on he le ) and an OPPL sc ip (do ed squa e) as inpu s, and pe o ms
he changes defined by he OPPL sc ip on he inpu on ology, he eby gene a ing a new ou pu on ology (modified on ology, on he igh ).
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 4 o 16
h p://www.jbiomedsem.com/con en /4/1/2
Figu e 3 OPPL-Galaxy a chi ec u e. The inne ci cle ep esen s he OPPL w appe and he ou e one Galaxy. Galaxy manages he da a and
pa ame e s ha will be passed o he OPPL w appe . In o de o pass, o ins ance, an on ology o he OPPL w appe , he on ology mus be fi s
uploaded o Galaxy (o passed o i om he ou pu o ano he Galaxy ool). Also, Galaxy manages he ou pu o he OPPL w appe : i can be
edi ec ed o o he Galaxy ools o downloaded and sa ed as a s andalone file. The OPPL w appe coo dina es he OPPL API ( o pa se he OPPL
sc ip and execu e i ), he OWL API ( o ead/w i e on ologies om s din/ o s dou and pe o m changes), and he chosen easone ( o pe o m
in e ences).
elies on he OWL API o loading and sa ing
on ologies, i can load on ologies in he ollowing
o ma s: OBOF [30], OWL (RDF/XML, OWL/XML,
Func ional OWL Syn ax, MOS), Tu le, and KRSS.
•OPPL sc ip : a fla file con aining he OPPL sc ip
ha , when execu ed, will pe o m he desi ed
changes in he a ge on ology. This file may be
c ea ed by using he P o ´
eg´
e OPPL plugin
ia
he
OPPL ex edi o (wi h au ocomple ion), he OPPL
sc ip builde , o he OPPL mac os ab (see he OPPL
manual [31] o de ails on how o c ea e OPPL
sc ip s).
•Ou pu o ma : he o ma o he ou pu on ology,
ei he OBOF o OWL (RDF/XML).
•Choose easone : he easone o be used o
pe o ming he in e ence, Pelle [32], He miT [33],
FaCT++ [34], o Elk [35].
The ou pu on ology can be eused as inpu o o he
Galaxy ools like ONTO- oolki [36], o downloaded om
he Galaxy Web in e ace so ha i can be used ou side
Galaxy, o example wi h P o ´
eg´
e o OBO-Edi [37].
OPPL-Galaxy includes a ious modules wi h di e se
unc ionali y, apa om execu ing OPPL sc ip s. Addi-
ionally, o he ools a e exploi ed as pa o he use cases
(NCBO-Galaxy [38], SPARQL-Galaxy, GO::Te mFinde ).
See Table 1 o de ails.
Figu e 4 OPPL-Galaxy Web in e ace. The OPPL-Galaxy Web in e ace is displayed in he middle pane. In he le pane, a lis o s anda d Galaxy
ools is shown; in he igh pane, a sample o a his o y o he execu ed asks is shown.
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 5 o 16
h p://www.jbiomedsem.com/con en /4/1/2
Table 1 OPPL-Galaxy dis ibu ion and ela ed Galaxy ools
OPPL-Galaxy bundle
OPPL Execu es OPPL sc ip s
OWL Que y Pe o m DL (Desc ip ion Logics) que ies agains OWL on ologies, e u ning a lis o named
en i ies ha sa is y he que y
OPPL Que y Pe o m OPPL que ies, hus, que ies ha mix MOS wi h a iables
In e ence Add he in e ed axioms o he inpu on ology as asse ed axioms, gene a ing a new on ology
ha includes all he axioms
Me ge Resol es he impo axioms and adds he impo ed on ology o he inpu on ology file
NCBO-Galaxy bundle
The NCBO-Galaxy bundle includes modules o e ie ing on ologies, ex ac ing sub ees om on ologies,
sea ch o e ms in on ologies, anno a e ex s agains on ologies, e c. using NCBO Web se ices. See [38] o de ails
SPARQL-Galaxy bundle
SPARQL-Galaxy includes a ool o pe o ming SPARQL que ies on an OWL (RDF/XML) on ology;
i can be downloaded om he Galaxy Tool Shed (h p:// oolshed.g2.bx.psu.edu), unde ‘On ology manipula ion’.
Galaxy-OBO
Galaxy-OBO [39] is a o k o Galaxy ha includes w appe s o common ools like GO::Te mFinde [40]
This able p o ides a de ailed lis o he OPPL-Galaxy ools and o he ools ha a e execu ed in he wo kflows o he use cases.
Resul s
This sec ion p o ides use cases no only demons a ing
he u ili y o OPPL-Galaxy bu also showing, h ough
examples, how o use i . The use cases a e desc ibed in
de ail in [41]. All he use cases a e p o ided as Galaxy
wo kflows o use s o be able o execu e hem wi hou
ha ing o ebuild he use case om sc a ch. The URLs o
he wo kflows a e summa ised a Table 2.
Basic usage
The OPPL-Galaxy bundle includes a simple OPPL sc ip
o es ing pu poses ha wo ks wi h he es on ology
also included in he bundle (Figu e 5). The OPPL sc ip
is desc ibed as ollows o help he eade unde s and he
emainde o he use cases (mo e OPPL examples can be
ound a he OPPL sc ip s collec ion [42]):
1 ?agen :CLASS,
2 ?p ocess:CLASS
3 SELECT ?agen SubClassO pa icipa es _in
some ?p ocess
4 WHERE ?agen != No hing
5 BEGIN
6 ADD ?agen SubClassO pa icipa es _in
only ?p ocess
7 END;
Table 2 Galaxy wo kflows o ep oducing he use cases
Name Galaxy wo kflow
Basic usage h p://bio d .o g:8090/u/mikel-egana-a angu en/w/basic-usage-1
On ology debugging and
e alua ion∗
h p://bio d .o g:8090/u/mikel-egana-a angu en/w/on ology-debugging-and-e alua ion
Complex que ying o GO h p://bio d .o g:8090/u/mikel-egana-a angu en/w/complex-que ying-o -go
Expansion o gene p oduc
anno a ions h ough GO
s uc u e
h p://bio d .o g:8090/u/mikel-egana-a angu en/w/expansion-o -gene-p oduc -anno a ions- h ough-go-s uc u e
Selec i e ex ac ion o modules
om GO o e m en ichmen
h p://bio d .o g:8090/u/mikel-egana-a angu en/w/selec i e-ex ac ion-o -modules- om-go- o - e m-en ichmen
OWL TBox o ABox
ans o ma ion o assis ing
SPARQL que ies
h p://bio d .o g:8090/u/mikel-egana-a angu en/w/owl- box- o-abox- ans o ma ion- o -assis ing-spa ql-que ies
The name o he use case (as pe sec ion name) is p o ided in he le column; he URL o he Galaxy wo kflow is p o ided in he igh column. In o de o execu e a
wo kflow, he da ase s (on ologies, OPPL sc ip s, GAFs, e c.) mus be aken om he his o y (h p://bio d .o g:8090/u/mikel-egana-a angu en/h/oppl-galaxy-use-
cases- o -jbs) o he wo kflow can be ep oduced manually wi h he same da ase s, by uploading hem. The wo kflow “On ology debugging and e alua ion” ob ains
he on ologies di ec ly om NCBO se ices. Fo de ailed ins uc ions, see h p://wilkinsonlab.in o/OPPL-Galaxy. All he wo kflows can be ep oduced in a local Galaxy
ins alla ion; in o de o do so, he wo kflows and da ase s can be downloaded om h p://bio d .o g:8080/JBSusecases. a .gz.
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 6 o 16
h p://www.jbiomedsem.com/con en /4/1/2
Figu e 5 Basic usage. The OPPL engine akes he a ge on ology and OPPL sc ip as inpu s, and gene a es a new on ology changed acco ding o
he OPPL sc ip . The OPPL sc ip que ies he easone o a class wi h a ce ain es ic ion (SELECT ... WHERE clause, blue) and adds ano he
es ic ion o he e ie ed class (ADD clause, ed).
Lines 1 and 2 show he decla a ion o wo a iables
(?p ocess and ?agen )and hei ype(CLASS). These
a iables ep esen (se s o ) OWL classes. Then, line 3
in oduces a SELECT clause, which is p ocessed by OPPL
and sen o he easone asking o he classes ha a e sub-
classes o he anonymous exp ession pa icipa es in
some ?p ocess: he exp ession is w i en in MOS
and i mixes named en i ies o he on ology ( he p op-
e y pa icipa es in)wi h a iables(?p ocess and
?agen , ep esen ing se s o classes). La e , in line 4,
he classes e ie ed as membe s o he a iable ?agen
a e checked o sa isfiabili y (?agen != No hing).
Finally, he axiom SubClassO pa icipa es in
only ?p ocess is added (ADD) o he inpu on ology,
esol ing ?agen and ?p ocess o all he classes ha
ha e been bound and combina ions he eo .
On ology debugging and e alua ion
On ology debugging ( he p ocess o fixing de ec s in an
on ology) can be a daun ing ac i i y, especially when he
on ology he scien is is wo king wi h has no been de el-
oped in-house and/o i i p esen s a complex axioma i-
sa ion o e many en i ies. OPPL-Galaxy can be used o
de ec ing and fixing ce ain s uc u es ha a e consid-
e ed bad p ac ice (an ipa e ns) o a leas ‘suspicious’. The
de ec ion o an ipa e ns also offe s a ‘pic u e’ o he on ol-
ogy: i can be used o e alua e he o e all s uc u e o
he on ology as one o he c i e ia o judge i s quali y.
OPPL-Galaxy p o ides a means o defining an ipa e ns as
‘ es uni s’ ha can be un au oma ically agains a se o
on ologies, as pa o Galaxy wo kflows.
The no ion o an ipa e ns in on ologies has al eady
been in oduced [43,44]. Fo example, [44] men ions
using he OWL uni e sal es ic ion (only) wi hou any
o he es ic ion on he same p ope y (e.g. some)asa
po en ial an ipa e n (exclusi e uni e sal). This is due o
he ac ha , he only es ic ion, on i s own, can be
i ially sa isfied by an unsa isfiable (emp y) class, e.g. A
subclasso p only (B and C) can be sa isfiable
e en when B disjoin Wi h C, since he seman ics o
only s a e ha i he e is a ela ion, i mus be o (B and
C),o none:(B and C) is emp y and he e o e would
sa is y he none case.
The exclusi e uni e sal s uc u e can be easily de ec ed
in, o example, BioPAX [45], by he ollowing OPPL sc ip
(Figu e 6):
1 ? a ge :CLASS,
2 ?p op:OBJECTPROPERTY,
3 ? ille :CLASS
4 SELECT ASSERTED ? a ge SubClassO ?p op
only ? ille
5 WHERE FAIL ? a ge SubClassO ?p op some
? ille
6 BEGIN
7 ADD ? a ge SubClassO !OnlyBadP ac ice-
Resul
8 END;
This sc ip de ec s he exclusi e uni e sal s uc u eb
and adds all he classes ha p esen i as subclasses o
OnlyBadP ac iceResul ,aclassc ea edon hefly
i i does no exis in he on ology (!symbol). No e he
use o he ASSERTED keywo d (only he asse ed axioms,
no he in e ed ones, a e aken in o accoun : he ea-
sone is deac i a ed o que ying in o de o imp o e
pe o mance) and he FAIL keywo d (nega ion as ailu e,
which is ou o OWL seman ics, is used o de ec absen
exis en ial es ic ions).
The on ology can also be simply que ied, wi hou mod-
i ying i , by using he OPPL-Que y ool (See Table 1 and
Figu e 7):
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 7 o 16
h p://www.jbiomedsem.com/con en /4/1/2
Figu e 6 On ology debugging and e alua ion. This sc ip de ec s any class ha has a uni e sal es ic ion wi hou an exis en ial es ic ion
(do ed blue line). I adds a subClassO OnlyBadP ac iceResul axiom ( ed a ow) o any ma ching class.
1 ? a ge :CLASS,
2 ?p op:OBJECTPROPERTY,
3 ? ille :CLASS
4 SELECT ASSERTED ? a ge SubClassO ?p op
only ? ille
5 WHERE FAIL ? a ge SubClassO ?p op some
? ille
The exclusi e uni e sal s uc u e can also be modi-
fied by adding an exis en ial es ic ion o e e y uni e sal
es ic ion:
1 ? a ge :CLASS,
2 ?p op:OBJECTPROPERTY,
3 ? ille :CLASS
4 SELECT ASSERTED ? a ge SubClassO ?p op
only ? ille
5 WHERE FAIL ? a ge SubClassO ?p op some
? ille
6 BEGIN
7 ADD ? a ge SubClassO ?p op some
? ille
8 END;
E en hough he exclusi e uni e sal s uc u e migh
be conside ed as a legi ima e modelling decision, i is
ecommendable, o make su e he e is no i ially sa -
isfiable classes, o add exis en ial es ic ions on he fly
(and possibly o make en i ies disjoin ), apply eason-
ing o de ec i ially sa isfiable classes, and hen emo e
he exis en ial es ic ions again. Such p ocedu e can be
au oma ically pe o med using OPPL-Galaxy. An al e -
na i e would be o check he consis ency o he fille ,
e.g. ? ille subClassO owl:No hing, wi h he
easone ac i a ed, ins ead o checking o he exclusi e
uni e sal s uc u e [46].
Mo e an ipa e ns can be ound in he collec ion p e-
sen ed in [43]:
•Logical An ipa e ns (LAP): modelling e o s ha a e
de ec able by an au oma ed easone ,
e.g.
unsa isfiable
classes.
•Non-Logical An ipa e ns (NLAP): modelling e o s
ha a e no de ec able using a easone , usually
c ea ed by he de elope due o a misunde s anding
o he language seman ics ( he logical consequences
o he axioms s a ed in he on ology).
•Guidelines (G): al e na i e, simple axioma ic
exp essions o he same knowledge.
Synonym O Equi alence (SOE) is an example o a
NLAP. Such ype o an ipa e n desc ibes he si ua ion in
which wo classes a e decla ed as being equi alen and
bo h pe ain o he same on ology (i.e., heyha eno
been impo ed). Gene ally, ha means ha he de el-
ope in ends o model a synonym, which should be an
d s:label s ing, as a whole class. Such s uc u e
can be easily de ec ed, o example, in he NIF G oss
Ana omy on ology [47], using he ollowing sc ip (which
also emo es he non-desi ed s uc u e):
1 ? a ge :CLASS,
2 ? ille :CLASS
3 SELECT ASSERTED ? a ge equi alen To
? ille
4 BEGIN
5 REMOVE ? a ge equi alen To ? ille
6 END;
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 8 o 16
h p://www.jbiomedsem.com/con en /4/1/2
Figu e 7 OPPL que y ool. Web in e ace o he OPPL que y ool.
We do no claim ha hese s uc u es (exclusi e uni-
e sal in BioPAX and SOE in NIF G oss Ana omy) a e
e oneous pe se. We a he s a e ha , acco ding o he
expe ience o he au ho s o [43,44], and ou s, hey a e
modelling p ac ices ha may yield unexpec ed esul s
when au oma ed easoning is applied downs eam. The e-
o e, a scien is who migh euse hose on ologies should
be awa e o he exis ence o he men ioned an ipa e ns.
OPPL-Galaxy is a s aigh o wa d, powe ul and flexi-
ble ool ode ec an ipa e nsen masse when execu ed
as a Galaxy wo kflow: a scien is can ha e a collec ion o
an ipa e ns o he choice eady o be applied in any on ol-
ogy she wan s o euse (any an ipa e n can be defined by
he , since OPPL is, oughly, a supe se o OWL). The ull
p ocess can be au oma ed, defining once wha on ologies
o ob ain and hen adding an ipa e ns o he collec ion
as needed. Once he wo kflow has been execu ed and
he an ipa e ns de ec ed in he a ge on ology, she can
decide i he on ology mee s he equi emen s. Addi ion-
ally, OPPL-Galaxy can be used o modi y he on ologies
ha do no mee he equi emen s, wi hin he same wo k-
flow.
Complex que ying o GO
OPPL-Galaxy can be combined wi h o he Galaxy-
enabled ools o build ad anced wo kflows such as he one
shown in Figu es 8 and 9. This wo kflow can be used by a
scien is o pose a complex ques ion agains GO, namely
‘Wha a e he p o eins ha ac on p ocesses ha in ol e
hepa ocy es and a e pa o o egula e o he biological
p ocesses?’. Posing such a complex ques ion equi es di -
e en s eps ha can be pe o med wi h OPPL and s o ed
o u he analysis wi h he help o Galaxy.
The wo kflow execu es he OPPL que y ool and he
Galaxy ool o compa ing wo da a se s (included in
he s anda d Galaxy dis ibu ion, in ‘Join, sub ac and
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 9 o 16
h p://www.jbiomedsem.com/con en /4/1/2
Figu e 8 Complex que ying o GO (as shown in Galaxy). OPPL-que y wo kflow o que ing GO agains GAFs. The esul is a lis o p o eins o
in e es .
g oup’). Thus, his wo kflow combines Galaxy ools o
e ie e exac ly he p o eins ha he scien is defined in
he plain-english que y, which is ansla ed in o a machine
in e p e able o m, as discussed below.
The OPPL sc ip que ies GO o he e ms ha ha e
‘Hepa ocy e’ as pa o hei names and ha a e ela ed,
ia pa o o egula es, o a biological p ocess:
1 ?hepa ocy e _p ocess:CLASS,
2 ?hepa ocy e _p ocess _label:CONSTANT =
MATCH(".?hepa ocy e.+"),
3 ?pa _o _o _ egula es:OBJECTPROPERTY
4 SELECT ?hepa ocy e _p ocess.IRI label
?hepa ocy e _p ocess _label,
5 ?hepa ocy e _p ocess subClassO
?pa _o _o _ egula es some
GO _0008150
Then, he Galaxy ool o compa ing wo da a se s is
used o ex ac he p o eins in ol ed in he esul ing p o-
cesses o in e es , using he GO e ms as keys agains
a Gene Associa ion File (GAF) [48]. The esul o his
compa ison is a lis o he p o ein iden ified as o in e es .
This wo kflow demons a es some o he main ad an-
ages p o ided by OPPL-Galaxy: on one hand, his ype
o analysis can only be pe o med, effec i ely, wi h OPPL
(see below). On he o he hand, he unique capabili ies o
OPPL a e enhanced due o he ac ha hey a e execu ed
Figu e 9 Complex que ying o GO (de ails). De ailed depic ion o he wo kflow shown in Figu e 8.
Ega˜
na A angu en e al. Jou nal o Biomedical Seman ics 2013, 4:2 Page 16 o 16
h p://www.jbiomedsem.com/con en /4/1/2
14. Leonelli S, Diehl A, Ch is ie K, Ha is M, Lomax J: How he gene on ology
e ol es. BMC Bioin 2011, 12:325+.
15. Luigi Iannone: OPPL. [h p://oppl.s .ne ]. [Online; accessed
28-Ma ch-2012]
16. ˇ
S ´
ab Zamazal O, S ´
a ek V, Iannone L: Pa e n-based on ology
ans o ma ion se ice exploi ing OPPL and OWL-API. In EKAW’10;
2010:105–119.
17. Ega˜
na M, Rec o A, S e ens R, An ezana E, Gangemi A, Euzena J:
Applying On ology Design Pa e ns in Bio-on ologies. In EKAW 2008,
LNCS 5268; 2008:7–16.
18. Fe nandez-B eis JT, Iannone L, Palmisano I, Rec o AL, S e ens R:
En iching he Gene On ology ia he dissec ion o labels using he
On ology P e P ocesso Language. In EKAW; 2010:59–73.
19. Iannone L, Palmisano I, Rec o AL, S e ens R: Assessing he Sa e y o
Knowledge Pa e ns in OWL On ologies. In ESWC; 2010:137–151.
20. Iannone L, Rec o A, S e ens R: Embedding Knowledge Pa e ns in o
OWL. In ESWC; 2009:218–232.
21. Mo ensen JM, Ho idge M, Musen MA, Noy NF: Modes use o
On ology Design Pa e ns in a eposi o y o biomedical on ologies.
In WOP. Edi ed by Gangemi A, Blomq is E:2012.
22. Jupp S, Ho idge M, Iannone L, Klein J, Owen S, Schans a J, Wols enc o
K, S e ens R: Populous: a ool o building OWL on ologies om
empla es. BMC Bioin o ma ics 2012, 13(Suppl 1):S5+.
23. Goecks J, Nek u enko A, Taylo J, Galaxy Team: Galaxy: a comp ehensi e
app oach o suppo ing accessible, ep oducible, and anspa en
compu a ional esea ch in he li e sciences. Genome Biol 2010,
11(8):R86+.
24. Ho idge M, D ummond N, Goodwin J, Rec o AL, S e ens R, Wang H: The
Manches e OWL Syn ax. In OWLED, Volume 216 o CEUR Wo kshop
P oceedings. Edi ed by G au BC, Hi zle P, Shankey C, Wallace E, G au BC,
Hi zle P, Shankey C, Wallace E; 2006.
25. Goble CA, Bhaga J, Alekseje s S, C uickshank D, Michaelides D, Newman
D, Bo kum M, Bechho e S, Roos M, Li P, De Rou e D: myExpe imen : a
eposi o y and social ne wo k o he sha ing o bioin o ma ics
wo kflows. Nucleic Acids Res 2010, 38(suppl 2):W677—W682.
26. Galaxy p ojec : Galaxy Tool XML File. [h p://wiki.g2.bx.psu.edu/Admin/
Tools/ToolConfigSyn ax]. [Online; accessed 28-Ma ch-2012]
27. S an o d Cen e o Biomedical In o ma ics Resea ch: P o ´
eg´
e. [h p://
p o ege.s an o d.edu/]. [Online; accessed 28-Ma ch-2012]
28. Luigi Iannone: OPPL API. h p://sou ce o ge.ne /p ojec s/oppl2/files/
OPPLAPI/. [Online; accessed 28-Ma ch-2012]
29. Ma hew Ho idge: OWL API. h p://owlapi.s .ne . [Online; accessed
28-Ma ch-2012]
30. Mungall C, Ru enbe g A, Ho ocks I, Osumi-Su he land D, An ezana E,
Balhoff J, Cou o M, Die ze H, Day-Rich e J, I eland A, Lewis S, Manzoo S,
Ti mizi S H: OBO Fla File Fo ma Syn ax and Seman ics. h p://pu l.
obolib a y.o g/obo/obo o ma /spec.h ml. [Online; accessed
28-Ma ch-2012]
31. Mikel Ega˜
na A angu en: OPPL manual. h p://oppl2.sou ce o ge.ne /
manual.pd . [Online; accessed 28-Ma ch-2012]
32. Cla ck and Pa sia LLC: Pelle : OWL 2 Reasone o Ja a. h p://
cla kpa sia.com/pelle /. [Online; accessed 28-Ma ch-2012]
33. In o ma ion Sys ems G oup: He mi OWL Reasone . h p://www.he mi -
easone .com/. [Online; accessed 28-Ma ch-2012]
34. Dmi y Tsa ko : FaCT++. h p://code.google.com/p/ ac plusplus/.
[Online; accessed 28-Ma ch-2012]
35. Ye geny Kazako : Elk easone . h p://code.google.com/p/elk-
easone /. [Online; accessed 22-Oc obe -2012]
36. An ezana E, Venka esan A, Mungall C, Mi ono V, Kuipe M:
ONTO-ToolKi : enabling bio-on ology enginee ing ia Galaxy. BMC
Bioin o ma ics 2010, Suppl 12:S8+.
37. Day-Rich e J, Ha is MA, Haendel M: Gene On ology OBO-Edi Wo king
G oup, Lewis S. OBO-Edi –an on ology edi o o biologis s.
Bioin o ma ics 2007, 23(16):2198–200.
38. Jos´
e An onio Mi˜
na o-Gimenez, Mikel Ega˜
na A angu en, Jesualdo Tom´
as
Fe n´
andez-B eis, E ick An ezana: NCBO-Galaxy: b idging he BioPo al
web se ices and he Galaxy pla o m. In ICBO; 2012. [So wa e demo].
39. Ch is Mungall: galaxy-obo. h ps://bi bucke .o g/cmungall/galaxy-obo/
o e iew. [Online; accessed 22-Oc obe -2012]
40. Boyle EI, Weng S, Gollub J, Jin H, Bo s ein D, Che y JM, She lock G:
GO::Te mFinde -open sou ce so wa e o accessing Gene On ology
in o ma ion and finding significan ly en iched Gene On ology
e ms associa ed wi h a lis o genes. Bioin o ma ics,20(18):3710+.
41. Mikel Ega˜
na A angu en, E ick An ezana, Jesualdo Tom´
as Fe n´
andez-B eis:
OPPL Galaxy. h p://wilkinsonlab.in o/OPPL-Galaxy. [Online; accessed
20-Oc obe -2012]
42. Luigi Iannone: OPPL Examples. h p://oppl2.sou ce o ge.ne /
aggedexamples/. [Online; accessed 28-Ma ch-2012]
43. Co cho O, Roussey C, Vilches Bl´
azquez LM, P´
e ez IE a Blomq is FSVS, Ku
Sandkuhl (Eds): Pa e n-based OWL On ology Debugging Guidelines; WOP
2009.
44. Rec o A, D ummond N, Ho idge M, Roge s J, Knublauch H, S e ens R,
Wang H, W oe C: OWL Pizzas: P ac ical Expe ience o Teaching
OWL-DL: Common E o s and Common Pa e ns. In Enginee ing
Knowledge in he Age o he Seman icWeb, Volume LNAI 3257. Edi ed by
Mo a E, Shadbol N, S u A, Gibbins N; 2004:63–81.
45. BioPAX g oup: BioPAX Le el 3 on ology. h p://www.biopax.o g/
elease/biopax-le el3.owl. [Online; accessed 28-Ma ch-2012]
46. Anonymous e iewe : Commen . SWAT4LS manusc ip e iew 2011.
47. Bug WJ, Ascoli GA, G e he JS, Gup a A, Fennema-No es ine C, Lai d AR,
La son SD, Rubin D, Shephe d GM, Tu ne JA, Ma one ME: The NIFSTD
and BIRNLex ocabula ies: building comp ehensi e on ologies o
neu oscience. Neu oin o ma ics 2008, 6(3):175–94.
48. Camon E, Mag ane M, Ba ell D, Lee V, Dimme E, Maslen J, Binns D, Ha e
N, Lopez R, Apweile R: The Gene On ology Anno a ion (GOA)
Da abase: sha ing knowledge in Unip o wi h Gene On ology.
Nucleic Acids Res 2004, 32:D262—D266.
49. Jupp S, S e ens R, Hoehndo R: Logical Gene On ology Anno a ions
(GOAL): explo ing gene on ology anno a ions wi h OWL. J Biomed
Seman ics 2012, 3Suppl 1.
50. Heiko Die ze: Obo On ology Release Tool. h p://code.google.com/p/
owl ools/wiki/Oo In o. [Online; accessed 22-Oc obe -2012]
51. Gene On ology Conso ium: On ology p ocessing ecipes. h p://wiki.
geneon ology.o g/index.php/On ology p ocessing ecipes. [Online;
accessed 22-Oc obe -2012]
52. Mi˜
na o-Gimenez JA, Ega˜
na A angu en M, B´
eja RM, Fe n´
andez-B eis JT,
Mad id M: Seman ic in eg a ion o in o ma ion abou o hologs and
diseases: The OGO sys em. J Biomed In 2011, 44:1020–1031.
53. W3C: OWL 2 Web On ology Language New Fea u es and Ra ionale.
h p://www.w3.o g/TR/owl2-new- ea u es/. [Online; accessed
28-Ma ch-2012]
54. Vassiliadis V, Wielemake J, Mungall C: P ocessing OWL2 on ologies
using Thea: An applica ion o logic p og amming. In OWLED; 2009.
55. OWLlink Wo king G oup: OWLlink. [h p://www.owllink.o g/. [Online;
accessed 28-Ma ch-2012]]
56. Race Sys ems GmbH and Co KG: Race P o 2.0. h p://www. ace -
sys ems.com/. [Online; accessed 28-Ma ch-2012]
57. Blond´
e W, Mi ono V, Venka esan A, An ezana E, De Bae s B, Kuipe M:
Reasoning wi h bio-on ologies: using ela ional closu e ules o
enable p ac ical que ying. Bioin o ma ics 2011, 27(11):1562–1568.
58. Galaxy p ojec : Galaxy on he cloud. h p://wiki.g2.bx.psu.edu/Admin/
Cloud. [Online; accessed 28-Ma ch-2012]
59. Diehl AD, Augus ine AD, Blake JA, Cowell LG, Gold ES, Gond ´
e-Lewis TA,
Masci AM, Meehan TF, Mo el PA, Nijnik A, Pe e s B, Pulend an B,
Scheue mann RH, Yao QA, Zand MS, Mungall C J: Hema opoie ic cell
ypes: P o o ype o a e ised cell on ology. J Biomed In 2011,
44:75–79.
60. Mikel Ega˜
na A angu en, W oe C, Goble C, S e ens R: In si u mig a ion o
handc a ed on ologies o eason-able o ms. Da a and Knowledge
Enginee ing 2008, 66:147–162.
61. Mik oyannidi E, Rec o A, S e ens R: Abs ac ing and Gene alising he
Founda ional Model Ana omy (FMA) On ology. In Bio-On ologies;
2009.
doi:10.1186/2041-1480-4-2
Ci e his a icle as: Ega˜
na A angu en e al.:OPPL-Galaxy, a Galaxy ool
o enhancing on ology exploi a ion as pa o bioin o ma ics wo kflows.
Jou nal o Biomedical Seman ics 2013 4:2.
